Prebiopsy Magnetic Resonance Imaging and Prostate Cancer Detection: Comparison of Random and Targeted Biopsies

Abstract

We compared the accuracy of visual targeted biopsies vs computerized transrectal ultrasound-magnetic resonance imaging registration using a rigid (Esaote®, nondeformable) or elastic (Koelis®, deformable) approach. A total of 391 consecutive patients with suspected localized prostate cancer were prospectively included in analysis. All patients underwent prostate magnetic resonance imaging, followed by 10 to 12-core random prostate biopsies. When magnetic resonance imaging detected suspicious findings, targeted biopsy was performed, including visual, rigid system and elastic system targeted biopsies in the first 127 patients, the next 131 and the last 133, respectively. Cancer detection rates were assessed by conditional logistic regression. Targeted biopsies alone and random biopsies were further compared for the amount of tissue sampled and microfocal cancer detection, the latter defined as a single core with 5 mm or less of Gleason 6 cancer. Patient characteristics and random biopsy detection rates were similar among the groups. Magnetic resonance imaging detected at least 1 suspicious area in 54 (42%), 78 (59%) and 82 patients (62%) in groups 1, 2 and 3, respectively. The cancer detection rates of rigid and elastic system targeted biopsies were significantly higher than the random biopsy rate (p = 0.0065 and 0.0016, respectively). Visual targeted biopsy did not perform better than random biopsy (p = 0.66). Rigid and elastic system targeted biopsies allowed for decreasing the number of cores and the detection of microfocal cancer, while increasing the detection of high grade cancer. When performed with computerized magnetic resonance imaging-transrectal ultrasound image registration, targeted biopsy alone improved cancer detection over random biopsies, decreased the detection rate of microfocal cancer and increased the detection rate of cancer with a Gleason score of greater than 6.